Though the genome's sequence is essentially ?xed, its state may be constantly changing inside a cell. Two aspects of this changing state are the speci?c arrangement of myriad protein complexes along the genome in the form of chromatin, and the current rate of transcript production for each gene. A fundamental research objective is to understand the relationship between these two, and in particular, how transcript production rates (TxPRs) are in?uenced by genome-wide chromatin state. The goal of this proposal is to develop models that are capable of predicting a cell's genome-wide transcription state from knowledge of its genome-wide chromatin state. To build such models requires simultaneously pro?ling a cell's genome-wide chromatin state and transcription state at different times or under different conditions: Observing how the two change together, particularly in the context of directed perturbation, provides the statistical leverage needed to build predictive models that can provide causal insight. In this proposal, models will be developed and validated by monitoring chromatin and transcription in budding yeast as they progress through the cell cycle, a temporal series of highly regulated events controlling cell proliferation, aberrations of which can lead to cancer. Owing to the complexity of this challenge, yeast is used as a starting point because of its compact genome and genetic tractability, but we anticipate our methods will also be applicable in more complex organisms, including human. One major obstacle is that state-of-the-art chromatin immunoprecipitation (ChIP) methods for assaying chromatin state require a separate experiment not only for each time point and experimental condition, but also for each of the 100s-1000s of types of proteins binding along the genome. To overcome this hurdle, this proposal describes a novel method for ef?ciently learning quantitative genome-wide chromatin occupancy pro?les (GCOPs) using nuclease-digested chromatin at single-base resolution. The proposed method enables the comprehensive determination of quantitative chromatin occupancy of transcription factors, nucleosomes, and other DNA-binding factors across the entire genome without requiring a separate experiment for each. Producing GCOPs in conjunction with high- resolution measurements of TxPRs will allow the development of sophisticated, mechanistically interpretable models that predict transcript production rates as a function of chromatin state. The proposed research will result in (i) ef?cient new methods for producing quantitative GCOPs that will be applicable in any organism with a sequenced genome; (ii) GCOPs from budding yeast as they progress through the cell cycle, revealing for the ?rst time in any organism how genome-wide chromatin occupancy changes over the course of the cell cycle; (iii) characterization of how genome-wide chromatin changes are linked to changes in TxPRs, not only in wild-type yeast, but also under a wide range of genetic and genomic perturbations; and (iv) models learned from all these data that can predict TxPRs on the basis of chromatin occupancy, providing mechanistic insight into how the cell-cycle-regulated transcription program is in?uenced by its changing chromatin state.